Skip to Main content Skip to Navigation
New interface
Conference papers

A Convex Surrogate Operator for General Non-Modular Loss Functions

Abstract : Empirical risk minimization frequently employs convex surrogates to underlying discrete loss functions in order to achieve computational tractability during optimization. However, classical convex surrogates can only tightly bound modular loss functions, sub-modular functions or supermodular functions separately while maintaining polynomial time computation. In this work, a novel generic convex surrogate for general non-modular loss functions is introduced, which provides for the first time a tractable solution for loss functions that are neither super-modular nor submodular. This convex surro-gate is based on a submodular-supermodular decomposition for which the existence and uniqueness is proven in this paper. It takes the sum of two convex surrogates that separately bound the supermodular component and the submodular component using slack-rescaling and the Lovász hinge, respectively. It is further proven that this surrogate is convex , piecewise linear, an extension of the loss function, and for which subgradient computation is polynomial time. Empirical results are reported on a non-submodular loss based on the Sørensen-Dice difference function, and a real-world face track dataset with tens of thousands of frames, demonstrating the improved performance, efficiency, and scalabil-ity of the novel convex surrogate.
Document type :
Conference papers
Complete list of metadata
Contributor : Jiaqian Yu Connect in order to contact the contributor
Submitted on : Tuesday, April 12, 2016 - 2:01:57 PM
Last modification on : Thursday, February 3, 2022 - 3:01:40 AM
Long-term archiving on: : Wednesday, July 13, 2016 - 10:29:33 AM


Files produced by the author(s)


  • HAL Id : hal-01299519, version 1
  • ARXIV : 1604.03373


Jiaqian Yu, Matthew Blaschko. A Convex Surrogate Operator for General Non-Modular Loss Functions. The 19th International Conference on Artificial Intelligence and Statistics, May 2016, Cadiz, Spain. ⟨hal-01299519⟩



Record views


Files downloads